16

While studying for the CCENT exam, my reference materials have made an alarming number of references to class A/B/C networks. Thankfully they just treat Class A/B/C as shorthand for /8, /16, and /24 CIDR subnets, and don't make any mention of an implicit subnet from the first nibble. Still, it throws me off to have "Class B" pop up in a question or explanation and have to remind myself at every step that there is an implied /16 mask in there.

Is this a convention that is still widely used despite being obsoleted over two decades ago? Am I going to just have to get used to this from my senior admins? And, perhaps most importantly, does Cisco expect its certified technicians/associates/experts to accept and use classful network terminology? (Ignore the last question if it violates Cisco's exam confidentiality policy.)

Update: After switching to a more authoritative reference/study guide, it became clear that Cisco expects knowledge of actual classful networks, insofar as the official study guide dedicates several chapters to them. This makes the question less about the A/B/C terminology and more about if/why admins are expected to know about classful networks.

YouAreTheHat
  • 246
  • 1
  • 3
  • 8
  • In my data communications classes at school (which were all designed around Cisco Networking Academy courses), classful addressing was used as the stepping off point for learning everything else. There were classful addressing questions on some tests. No clue about beyond that, though. – canadmos May 15 '14 at 03:22
  • 1
    While most people today shun the concept of classful addressing, I personally think it's valuable to learn, both for a historic perspective and also because I think it gives you a good foundational understanding of IPv4 addressing and subnetting. I don't have a problem with someone learning it, teaching it or discussing it. – joeqwerty May 15 '14 at 04:05
  • 10
    The fact the CCENT uses A/B/C as shorthand for /8 /16 and /24 measurably diminishes the credibility of the certification in my book. Would you trust a pilot if the pilot's exam questions he answered used the word "handlebars" to refer to the control yoke or "railyard" to mean "airport"? Should you be allowed to write exam quiestions if you don't understand your own industry? – tylerl May 15 '14 at 05:49
  • Not to mention CCNA book and quizzes contain SO much errors you have to keep the errata list beside to verify each one. Unfortunately, there isn't any other certification in networks administration that is equally known by recruiters... – mveroone May 15 '14 at 07:41
  • 1
    IMO, any course or study material that doesn't teach classful addressing (at the very least from a historic perspective) is short changing you. I would have never understood CIDR and VLSM if I didn't learn and understand classful addressing first. - We use DNS instead of HOSTS files but that doesn't negate the need to know what the HOSTS file is, what it does, and how to use it if presented with a situation that requires it. The same can and should be said about classful addressing. – joeqwerty May 15 '14 at 17:20

5 Answers5

24

You should know three things about class-based-routing:

  • Class-based routing was a simpler system that was abandoned (in 1993) long before most people ever heard about the Internet. In all likelihood, nobody you will ever know has used it. And if any of your network equipment is that old, you should seriously consider alternate employment. The system used the first few bits of the address to determine its class, and (indirectly) its netmask. Note that the netmask was implied in the class, it did not determine the class. Saying you have a "Class C at 172.16.1.0" will earn you a swift kick from anyone with even a vague understanding of class-based routing.

  • People currently say Class A, B, and C to mean /8, /16, and /24 netmasks, respectively. As should be obvious from the above, they do so incorrectly. They typically think it makes them appear knowledgeable and wise to the history of of the Internet (oh, the irony).

  • Some hold-overs of the original system still exist. "Class D" (prefix 224 to 239) is still multicast, and "Class E" (prefix 240 to 255) is still "Reserved" or "Experimental". Plus, some (older) systems assume a default netmask based on the original class designation; so /8 for prefix 0 through 128, for example. This is often more annoying than helpful, but that's where it came from.

tylerl
  • 14,885
  • 7
  • 49
  • 71
  • 2
    Windows still does the last one (default mask). Its stupid because Windows tcp/ip is about the same age as CIDR – Mark Henderson May 15 '14 at 07:47
  • 2
    I wonder if saying "class A, B and C" to mean "/8, /16 and /24" is such a big problem. Perhaps it's just something that changed its meaning with time. In the same category, how many people use "[RJ45](http://en.wikipedia.org/wiki/8P8C_modular_connector#8P8C)" correctly? I can't think of any networking device with an 8P8C networking connector that doesn't say "RJ45" in its spec instead. – Bruno May 15 '14 at 21:48
  • "Class C" isn't any *more clear* than "/24". You have to learn /24 anyway, so calling it "Class C" just serves to confuse people who aren't familiar with the terminology while at the same time making yourself look uninformed to people who are. It's a bit like referring to the "OSI model of TLS encryption" -- that is, (A) most people don't even know what the OSI model is, and (B) apparently neither do you. – tylerl May 16 '14 at 00:39
  • 2
    I'm just saying it seems to have become a common expression (like everyone uses "RJ45" to talk about something that's not the RJ45 standard). Although it's incorrect, some people might find these class names easier to say than "/24", which in turns means than most people collectively know what they're referring to (I'm not endorsing the practice, I'm just saying it's not necessarily the end of the world). I'm not sure what you're saying about the OSI model and what I know about it or not... Where was I mistaken about the OSI model? – Bruno May 16 '14 at 00:54
  • I'm just saying that the terminology is popular with a very, very, very small subset of the industry. And completely coincidentally, they all seem to be Cisco Certified Engineers. – tylerl May 16 '14 at 00:57
  • And it seems oddly coincidental that the non-Cisco trained folks are the ones implying it doesn't need to be learned nor-understood yet in my experience these non-Cisco folk are the ones that usually break the network due to ignorance. – Chef Flambe May 16 '14 at 04:06
  • 2
    @chefflambe I have dealt with plenty of cisco folk who barely understood subnetting. There is plenty of ignorance to go around, regardless of what exams someone managed to pass. – Grant May 16 '14 at 14:21
23

The only people I have ever met who use the terms "Class A/B/C" are people who are using them incorrectly to mean a /8 or a /16 etc (except for just one person who used "Class A" to be the first octet, "Class B" to be the second octet, etc. But he was an idiot who refused to change his phrasing even though it led to so much confusion).

So go ahead and learn about classful addressing. More knowledge is better. But the most important thing to learn about "Class A/B/C" addresses is:

  1. You should never actually use those terms
  2. When you hear someone else using those terms, they are using them incorrectly

It's a good way of telling who actually knows their stuff and is experienced outside of theory. It would throw up red flags if I heard the terms used incorrectly in a job interview.

Mark Henderson
  • 68,316
  • 31
  • 175
  • 255
  • I should mention this is 100% correct, despite my answer below somewhat condoning its use in practice. – RobbieCrash May 15 '14 at 05:21
  • Are you implying that noone can use them correctly? – Angry Lettuce May 15 '14 at 15:53
  • 1
    @RagingScallion no, I'm saying that nobody I've met *does* use them correctly. I'm sure that there are people who could correctly say "We got our first Class C back in '89" (that would be correct). I've just never met them. So it's safe to assume that if you do hear it, it's being used wrong. If they're 45 or over and have a beard and only drink home-brews, then perhaps they are using it correctly. But chances are slim. – Mark Henderson May 15 '14 at 20:37
  • 1
    I'm almost 45, have a beard, enjoy home-brews. Yet I still (up til today) use(d) the terms incorrectly. +1 from me to OP and all answers like yours. – MattBianco May 16 '14 at 13:07
14

In a word, yes.

People will often talk about it in the same way that the course material does. Nobody uses classful networking, or at least nobody I've encountered; during infrastructure/architecture planning people will often say things like "we'll set them up with two class C's here." It's easier than saying "we'll give them a /23" or "172.16.0.0-172.16.1.254 should be enough" and it communicates the same message.

While it may not be technically 100% accurate, lots of the stuff that gets thrown around in industry conversation isn't to spec, eg: router/switch/AP\Bridges aren't referred to as anything other than routers.

Regarding the second question, I don't think it violates, but you're expected to understand the exam topics as it's presented in the material reading. That is, you need to know what they are, and you need to understand that they often refer to those specific subnets.

RobbieCrash
  • 1,131
  • 7
  • 25
0

No. It is not about IP network bitlength, and especially it isn't about which standard is already obsolete, and how long. It is about literacy. /A, /B, /C is used today by sysadmins who can't understand the binary arithmetic.

Cisco is making a lot of money by giving papers for people who never seen in their life a network before their costly Cisco "schooling". Effectively, it is like a "Polinese University". They don't know, what is an "ip network", they don't know what are "powers of two", or if they know, they can't use this knowledge.

But they are cheap, and want some type of paper. And Cisco gives these papers them, although it sells its market reputation by giving his certificates to these people.

If you hear from somebody: "Class-B Network", and not "16-bit network", or "64k network", or "/16 network", you know on the moment: this person probable can't say, who many different ips on a /22 network can concurrently exist.

It is this question is really about. Not about standards and not about old stories. This question is about technical literacy.

peterh
  • 4,914
  • 13
  • 29
  • 44
-1

I was working at this place which had hired an IT firm to configure some routers for the private network we had. The company I worked for usually had outside contractors do the grunt work.

Well the guy in our company in charge of the project decided on how he wanted to design the networks and was using a /24 mask. The problem was that he had no idea that routing protocols can be classful or classless. So this guy picks networks of 10.1.3.x / 10.1.4.x / 10.1.5.x and the such but choose to use a classful routing protocol which meant that nothing worked even though he set the masks for /24. Can't really fault him for not knowing as he didn't have any formal knowledge but he was such an arrogant ass about everything I had to laugh at him.

The IT firm, the "experts", were the ones that did the configurations and installation. For some strange reason they had no idea why it didn't work either. I walked up and took one look at the routing protocol and his network map and fixed this two week old problem in 5 minutes. The IT firm still charged for two weeks worth of "troubleshooting".

Little ole me with my CCENT certificate from 2007.

So yes, learning about Class based addressing is important because like it or not there is a crap load of legacy stuff out there and if you don't know what the major difference is between IGRP and OSPF or even the basic RIPv1 and RIPv2 and how the Class system applies to them you have no business taking people's money.

Take another look at those Cisco questions that mention ClassA/B/C. Unless they have a typo (okay 60% chance it will) they will be using examples that follow the Classful structure.

[Edit: I should mention this happened two years ago so its not some "old war story".]

Chef Flambe
  • 113
  • 5